How I Prepare

Samples for EBSD

Analysis

Matt Nowell
EBSD Product Manager
May 25, 2017

1
 Acknowledgements
‡ Ron Witt ± EBSD Analytical
‡ Stuart Wright, Rene de Kloe ± EDAX
‡ George Vander Voort ± Struers
‡ Lucille Giannuzzi - EXpressLO
‡ Joe Michael ± Sandia National Lab
‡ Allied High Tech
‡ Buehler
‡ Gatan
‡ E.A. Fischione Instruments
‡ FEI
‡ JEOL
‡ Hysitron
‡ 0DQ\PDQ\PRUHGLVFXVVLRQVDQGLGHDVZLWKSHRSOH«

2
 Background
‡ ,¶PQRWDVDPSOHSUHSDUDWLRQH[SHUW
± Use sample prep vendors as a resource
‡ I have looked at (and prepared after an initial look) a lot of EBSD
samples
‡ My analytical success often depends strongly on the quality of
the EBSD sample preparation
± EBSD does not work well if you do not get an EBSD pattern

‡ = Strong driving force for quality preparation

3
 Sample Prep as Black Magic

‡ Often sample preparation

can seen like more of an art
than a science
‡ Can make you feel like you
are on the outside looking in
LI\RXGRQ¶WNQRZWKHPDJLF
spell
‡ Over time, you do develop a
feel for what works

4
 The D in EBSD is Key

‡ EBSD is a diffraction-based
technique
± Riding the electron wave
‡ 7KH³EHWWHU´WKHFU\VWDOODWWLFHWKH
smaller the angular range where https://publish.illinois.edu/x-­‐raycrystallography/2014/12/18/introduction/

%UDJJ¶VODZLVVDWLVILHGDQGWKH
more intense the diffraction for a
given plane.
‡ Multiple planes diffract https://chem.libretexts.org/Core/Analytical_Chemistry/Analytical_Sciences_Digital_Library/JASDL/Coursew

simultaneously with EBSD are/Introduction_to_X-­‐ray_Diffraction_(XRD)

5
 Better Preparation = Sharper Patterns (Ideally)
‡ EBSD patterns from
ZrO2 sample polished
for SEM work (left) and
for EBSD work (right).
‡ EBSD polish improves
lattice quality at sample
surface and sharpens
diffraction bands.
‡ Want a representative
lattice on surface after
preparation
6
 There is a Great Disturbance in the Lattice

‡ Dislocations present in a Local  

misorientations
Local  
perturbations
material can disrupt the
periodicity of the local crystal
lattice.
‡ Can cause local lattice
curvature
‡ Can degrade EBSD pattern
quality

7
 Statistically Stored Dislocations (SSDs)

‡ Dislocations are present, but with a net

Burgers vector of zero
‡ Leads to degraded EBSD patterns
± Can detect with EBSD Image Quality
± IQ should be used carefully
‡ Effect depends on the interaction
volume size within lattice distortion
area
± SEM Voltage
± SEM Emitter
8
 Geometrically Necessary Dislocations (GNDs)
‡ GNDs create small changes in crystallographic orientation or
lattice curvature
‡ Interaction volume affects final EBSD pattern if sampling
distorted lattice region

9
 Why I Prepare EBSD Samples the Way I Do
‡ Typically I am showing what EBSD
can do for different samples,
materials, or applications.
± Want to make a good impression
‡ Sometimes comparing 2 or more
samples
± Want consistency for comparisons
‡ Look at a wide range of sample
types
± No typical sample
± 2IWHQGRQ¶WKDYHPDQ\GHWDLOVDERXW
sample origin or history
10
 My Goal ± 100% Indexing Success Rate
‡ §,65SRVVLEOHRQUDQJHRIPDWHULDOVDQGPDWHULDOVWDWHV
Rolled  Aluminum Ti6Al4V Deformed  Ferritic  Steel
99.8%  ISR 99.5%  ISR 99.9%  ISR

11
 What Happens if I am < 100% ISR?

‡ :KLOH,65LVQLFHLW¶VQRWDOZD\VDFKLHYDEOH
‡ ,W¶VEHQHILFLDOWRWU\DQGXQGHUVWDQGZK\FROOHFWHGSRLQWVFRXOG
not be indexed (or indexed correctly)
‡ This information may tell you more about your material and
microstructure
‡ It helps to know how the EBSD pattern indexing works
‡ It might also tell you your sample needs further/better
preparation

12
 Indexing at Grain Boundaries
‡ Near grain boundaries, the
interaction volume can
sample multiple
orientations causing
overlapping EBSD patterns
‡ Sometimes these can be
deconvoluted via indexing
approach
‡ Grain boundary
topography is also
common
13
 Indexing at Grain Boundaries
‡ In this case, we see a degradation of IQ and CI values at grain
boundary, but a sharp change in orientation (no noise)

14
 Crack Path

‡ Non-indexed points primary

show the crack path through
the microstructure of a
pipeline steel
‡ Can be used to determine the
grain boundary character of
the crack propagation
pathway

15
 Plastic Strain and Lattice Deformation

‡ Lattice strain introduced into

this Nitinol shape memory
alloy degraded EBSD pattern
quality
‡ The spatial shape distribution
of the non-indexed points
also suggest a 2nd phase
present
‡ IQ, PRIAS, and EDS info can
confirm this
16
 Other Phases

17
 Amorphous Regions

‡ In this example, a thermoelectric

material crystallizes within an
amorphous film matrix
‡ While the lack of indexing, or of a
visible diffraction pattern, does
not prove its amorphous, it can
be inferred from sample
processing history

18
 Porosity and Anti-Grains
‡ Non-indexed points can be grouped together spatially and
analyzed (termed anti-grains) as in this CdTe solar cell thin film

19
 Surface Topography
‡ Topography introduced by nanoindentation causes regions that
are blocked and causes bad indexing. Plastic strain visible in
both orientation (left) and IQ (right) maps.

20
 EBSD Camera Noise
Traditional NPAR

‡ EBSD pattern signal to noise

(SNR) levels can decrease
due to faster acquisition
speeds and/or lower beam
currents
‡ At some SNR threshold
(materials dependent), EBSD
pattern indexing will fail
‡ NPAR can help

21
NPAR ± Neighbor Pattern Averaging and Reindexing

A new approach to improving S/N while maintaining acquisition speeds

22
 NPAR for Reducing Camera Noise Effects
Traditional - 22% Indexing NPAR - 96% Indexing

23
 Is a 100% Indexing Goal for Everyone?
‡ Not necessarily
‡ Some threshold (90%, 95%, xxx) might be good enough
± Does it tell you what you want/need to know?
± Is any data lost or misrepresented?
‡ Will depend on the type of characterization or analysis is
necessary
‡ Good practice though is to use consistent preparation routines
when comparing samples
‡ Still good to try and understand why not reaching 100% ISR

24
 Evaluating Sample Preparation Methods

‡ EBSD provides a great tool for

evaluating the surface quality of a
given preparation procedure
‡ Have used approach to qualify
procedure for Inconel 600 EBSD
standard material

http://weirdsciencekids.com/thescientificmethod.html

25
 EBSD Patterns After Different Prep Steps

1200 Grit SiC 1ȝm Alumina 0.3ȝm Alumina 0.05ȝm CS 15 min

0.05ȝm CS 30 min 0.05ȝm CS 1 hr 0.05ȝm CS 2 hr 0.05ȝm CS 4 hr

Clearly there is an improvement from 1200 grit SiC to 1ȝm Alumina to 0.3ȝm and
beyond. It is difficult to evaluate the difference other than visually however.

26
 EBSD as a Tool to Evaluate Sample
Preparation
240 Grit SiC 1200 Grit SiC
‡ EBSD is an ideal tool
for measuring plastic
deformation on a small
scale
‡ Can detect the
deformation
introduced/remaining
after preparation
(down to detection KAM  Maps
Polished  Surface
limits)
27
 Evaluating Sample Preparation Methods
1

‡ This approach tells us 2 0.9

Image Quality
Fraction Indexed

hours of colloidal silica is

Kernel Misorientation

0.8

best 0.7

Normalized Values
‡ I rarely take this approach as 0.6

I look at a wide range of 0.5

materials and each 0.4

sample/material can behave

0.3

differently
0.2

0.1

‡ Will offer more opinions that I 0

KDYHQ¶WYHULILHGDVIDFW
1200 Grit ȝP 0 .3 ʅ m 0 .0 5 ʅ m C S 0 .0 5 ʅ m C S 0 .0 5 ʅ m C S 0 .0 5 ʅ m C S 0 .0 5 ʅ m C S
SiC Alu m in a Alu m in a 1 5 M in 3 0 M in 6 0 M in 1 2 0 M in 2 4 0 M in

28
 Evaluating Sample Preparation Methods
‡ I will explain what I do, and why I
do it
‡ Much of the reasons why are
based on opinion and theory, but I
have not systematically tested
‡ New tools and approaches are
DYDLODEOHEXW,KDYHQ¶WWULHGDOO
‡ Disclosure: I have not monetary
stake in any sample preparation
company. These are the products
I use. http://weirdsciencekids.com/thescientificmethod.html

29
 Mounting Sample

‡ I generally use a TechPress2 from

Allied High Tech
‡ Uses heat and pressure, so sample
must tolerate this
‡ I like this approach because its fast (<
10 minutes) and reproducible
‡ The more specific details I can share,
the easier to reproduce for others

30
 Mounting Compound
‡ I use the ProbeMet compound from Buehler
‡ I liked it when I first used it in 1998
‡ It is a Cu and SiO2 filled epoxy thermoset, with good edge
retention
± Feels lighter than other compounds with more Cu
‡ It is not perfectly conductive, but it does help
‡ *HQHUDOO\,XVH´PRXQWVEXWKDYHGRQH´DQG´
‡ I use instructions on the package
± 150º C, 290 bar pressure, 3 minute pre-heat, 1 minute-heat, 3 minute
cool with water

31
 Mounted Sample

‡ ´0RXQWVDOORZIRUHDVLHU
tilting to EBSD conditions
while examining more of the
sample area
‡ I try and measure the amount
of compound to use to get a
consistent sample height
± Will depend on height/volume
of sample

32
 Epoxy

‡ When a sample cannot tolerate thermoset resin, I use an epoxy

mix ± Generally I use Allied High Tech EpoxySet
‡ Hard, clear epoxy with low curing temperature (54ºC) and no
shrinkage
‡ §KRXUFXUHWLPH± can be reduced via heating
‡ Non-conductive, but I have mixed in Buehler Conductive Filler to
improve performance, but makes mount opaque

33
 Epoxy Samples

‡ Need to use spray to make

removing from mounting cups
easier
‡ Sometimes the back side of
the mount is rough, and can
be tricky to cut to get front and
back sides parallel
± Can influence prevision of tilt
values and beam position

34
 Grinding and Polishing

‡ I use an Allied High Tech MetPrep 3

with Power Head
‡ I used adhesive backed SiC ´GLVNV
‡ Can set individual or central force
± Was strong driving force in initial
purchase
± Often preparing single sample
‡ Can adjust time, force, and rotation
direction

35
 Grinding with Silicon Carbide (SiC) Papers

‡ ,XVH´DGKHVLYHSiC papers
± 240 grit
± 320 grit
± 400 grit
± 600 grit
± 800 grit
± 1200 grit
‡ My final SiC DEUDVLYHVL]HLV§
5µm

36
 SiC Papers are Consumables
‡ For each grit, I use 2 SiC papers for 30
seconds each
± Insures fresh and effective abrasive
± &DQIHHOLW¶VXVHGDIWHUVHFRQGVZLWKILQJHU
‡ I run with 10lbf and 150 RPM on the platen
± Can calculate pressure if necessary
‡ I use complementary rotation with the platen
and the head
± Helps keep front and back sides of sample
parallel
± Better for ComboScan larger area maps
37
 Grinding with SiC Papers
‡ I use a strong water flush to keep abrasive clean and sample
cool (if sample with tolerate ± if not Glycol)
‡ Will use more initial SiC papers if I see the entire surface has not
received an initial grind
± Can be issue with larger samples, or if sample plane not well aligned in
mount
‡ I like adhesive papers because they stick better than ring
retained clothes for me
‡ I clean head with water and paper towel between each grit size
to prevent contamination
38
 Polishing

‡ I use 1µm and 0.3µm Al2O3 as primary

abrasive
± I use water based suspensions
± More economical than diamond abrasives
‡ I use Imperial cloth from Allied High Tech as
primary polishing cloth
± All purpose cloth
± Low napped
± Synthetic Rayon

39
 Polishing with Al2O3

‡ For each abrasive size, I use 1 Imperial cloth for 10 minutes

‡ I run with 9 lbf and 130 RPM on the platen
± A little less force and a little less pressure than with SiC papers
‡ I use contra rotation between the platen and the wheel
± Removal rate is slower than with SiC papers
‡ Charge cloth with water, and then drip water through the run
‡ Liberally add abrasive during the run
‡ Clean head and platen with water after each abrasive

40
 Other Polishing Cloths and Abrasvies

41
 Other Polishing Cloths and Abrasvies
‡ This approach is a general approach
± Usually I try this on most materials
‡ Will try different cloths and abrasives when necessary
± ,I,¶PQRWKDSS\ZLWKP\UHVXOWVRULI,NQRZVDPSOHLVGLIILFXOW
± Hard materials
± Soft materials
± Multiphase materials (with different material removal rates)
± 7KLQPDWHULDOV GRQ¶WZDQWWRSROLVKWKURXJKWKHVDPSOH
± 5HDFWLYHPDWHULDOV GRQ¶WZDQWVDPSOHWRUHDFWZLWKZDWHUHWF
‡ Will ask sample prep vendors or read website descriptions and recipes for
ideas
± ,GRQ¶WFRPSDUHUHFLSHUHVXOWVYVVWDQGDUGUHVXOWV
42
 Vibratory Final Polishing

‡ Final polish is typically done with

vibratory polisher with colloidal silica
‡ Generally mounted sample is placed
in weighted holder to run on cloth
‡ Use Imperial cloth (same as polishing)
‡ Will polish anywhere from 15 minutes
to overnight, depending on sample
and timeframe

43
 Colloidal Silica
‡ Colloidal silica is a chemical-
mechanical polishing solution
‡ It can be messy, and there is
a 0.04µm version to help
there
‡ A water-free 0.05µm version
available for water-sensitive
materials
‡ Generally we use 0.02µm, as
it should give finest polish

44
 Vibratory Final Polishing
‡ The longer the polishing time, the better the surface but also the
greater chance of differential polishing rates with different
materials
‡ 2-4 hours typical
‡ Keep time constant for samples you want to comapre
‡ Keep the cloth clean with cover to prevent the solution from
drying and crystallizing to prolong lifetime
‡ Be aware of possible contamination from earlier samples
‡ Cleaning platen and making sure cloth is flat helps ensure the
sample will continue to move

45
 Mounting Samples for SEM
‡ We generally mount samples to
pin stubs for SEM work
‡ Hot glue gun provides fast and
strong adhesion with vacuum
compatibility
‡ For non-mounted sample, will
sometimes use silver paint
‡ Silver paint used to provide
grounding path
± Let silver paint dry

46
 Carbon Tape

‡ For mounting samples, I avoid carbon tape

‡ Carbo tape can creep over time, and cause the
sample to shift when tilted for EBSD work
‡ Carbon tape can be used to provide a ground
path.
‡ I typically use Copper tape for enhanced
conductivity

47
 Grounding the Sample
‡ ,W¶VLPSRUWDQWWRJURXQGWKH
sample to prevent charging
effects
‡ Generally I will paint a path from
sample to stub with silver paint
± Conductive mount not as
conductive as I would hope
‡ Sometimes I will run a piece of
copper tape from sample to
stage
‡ Can check with voltmeter
48
 Grounding the Sample

‡ Carbon coating can improve

conductivity
± I like using a thickness meter for
consistency and reproducability
‡ Low vacuum SEM
± <RXGRQ¶WQHHGPXFKWRUHGXFH
effects on tilted sample
‡ Image sample (semiconductor)
before mapping to reach a steady-
state condition
49
 NPAR on Non-Conductive Ceramic
20kV 5nA Beam Current 12kV 1.5nA Beam Current - NPAR

‡ Reducing electron done (both energy and current) can reduce

charging effects. NPAR can maintain indexing speeds.

50
 ,RQ(WFKLQJIRU³'LIILFXOW´0DWHULDOV

‡ I have used broad ion

beam polishing to
prepare materials like
magnesium and
zirconium
± Can remove oxide
layers easily
‡ Control angle and
energy of incident ion
beam
51
 Cross-Sectioning Ion Beam Polishers
‡ ,KDYHQ¶WXVHGD
ion CSP, but they
produce nice
EBSD patterns
‡ Very little ion
damage due to
beam geometry
‡ Some limitation in
ion polished area
Source: JEOL

52
 Focused Ion Beam
‡ FIB enables site-
specific EBSD
preparation on IQ

very small scale

TD IPF

FIB enables site-specific EBSD

preparation on a very small scale that ND IPF
Data courtesy of Joe Michael ±
would otherwise be impractical. Sandia National Laboratory

53
 Focused Ion Beam
milling strategy: milling strategy:

‡ Ion beam can remove material and grazing-incidence

edge-milling
low-incidence
surface-milling

leave an EBSD-quality surface

‡ Can control energy and incident
angle of ion beam
‡ Can use Ga or Xe ion beams
± Xe can prepare larger areas
‡ Can prepare samples in-situ large milling areas
feature has to be at required to avoid
± Used for Atom Probe Assist samples edge shadowing of EBSD

54
 Sample Preparation for EBSD

As-Deposited CdCl2 FIB Prepared CdCl2 FSD Orientation Contrast

Treated Treated Image

30kV 1nA Ga+ FIB (FEI Quanta 3D) used with glancing angle (1.5º) to cut
surface for EBSD collection

55
 Ion Beam Improving on Mechanical Polish
‡ In some cases (often multi-phase), ion beam polishing improves
EBSD results relative to mechanical polishing

Mechanical  Polishing Ion  Polishing

56
82.5%  ISR 98.4%  ISR
 Nanostructured Bi2Te3 Powder
0° Tilt 75° Tilt

Nanostructured Bi2Te3 powder produced by supersonic gas atomization as a method to reduce grain size

Samples  courtesy  of  and  in  collaboration  with  TXL  Group

57
 EBSD Patterns from Nanostructured Powder

58
 EBSD Maps from Polished Powder
Orientation  Map Grain  Map

:LWKQP(%6'VWHSVL]HJUDLQVL]H§QP
:LWKQP(%6'VWHSVL]HJUDLQVL]H§QP
59
 Summary

‡ Sample preparation is important in order to obtain quality,

representative EBSD patterns
‡ There are many different ways to prepare EBSD samples. For
each, the goal is the same
‡ At EDAX Draper, mechanical polishing is the primary method
used to prepare samples

‡ Thank you for your attention. Questions?

60
edax.com